HBI Deals+Insights / COVID-19

What’s behind Europe’s variation in excess death rates in 2022?

In this week’s infographics we look at monthly excess death rates across EU countries (plus Switzerland) throughout 2022. Surprisingly, the huge variation in excess death rates between countries is not explained by variations in the number of physicians or nurses per capita, nor in variations in health expenditure per capita, nor in variations in the proportion of private expenditure between countries.

Countries across Europe continued to see much higher than usual levels of excess deaths throughout 2022, despite Covid deaths being far lower than in 2020 and 2021. We looked at Eurostat’s monthly data on excess deaths, which are measured as a percentage rate of change in total deaths in a month compared to average monthly deaths in a baseline (pre-Covid) period (2016-2019). Eurostat has data going up to November 2022, so we looked at the figures for the first 11 months of 2022:

The excess death rate for the EU as a whole never went below 6% in 2022, and reached a peak of almost 17% in July.

But the figures for the EU as a whole mask wide differences between countries. So we calculated the average death rate across the 11 months for each country:

As you can see, there is a huge range. Sweden had an 11 month average excess death rate of just 2.6%, whilst Cyprus’ was close to 24%.

To find out what might be behind this variation we ran a multiple linear regression using a few possible explanatory variables as predictors of the 11 month average excess death rates. The predictor variables we used were the number of physicians per 1,000 residents, the number of nurses per 1,000 residents, the per capita expenditure (in euros) on health care, and the percentage of health care expenditure that is private. All of these vary significantly between the countries, and – assuming the excess deaths are caused by unusually high pressure on health care systems – would be obvious candidates for explaining why some countries are seeing higher excess death rates than others.

But, surprisingly, the multiple regression found no statistically significant correlations between any of the predictor variables and the 11 month average excess death rates, as you can see from the above scatter plots.

We would welcome your thoughts on this story. Email your views to Martin De Benito Gellner or call 0207 183 3779.